My colleagues and I had replaced a 20-ton, single-circuit condensing unit as part of a retrofit of a dual-duct variable-air-volume system. The cooling coil and drain pan in the air-handling unit had been replaced about a year earlier. After receiving a “too warm” service call, a technician found that the condensing unit on the low-pressure safety was off. The refrigerant charge was low. The next step was to find and repair the leak.

The cooling coil and connecting piping showed no leaks via an electronic “sniffer” or soap bubbles, and there was no sign of oil. However, there was evidence of a fine oil spray in the condensing unit's compressor compartment. Because electronic leak detectors tend to be ineffective on a roof in a breeze, out came the liquid soap. A tiny leak appeared at a valve cap, but that was it. R-22 refrigerant can leak pretty fast, and the port was in a reasonable place to explain the oil, so the valve and cap were replaced, the unit charged, the oil spray cleaned up, and the cooling restored. I reviewed the situation with the technician, and all seemed routine until the next call, which came about two weeks later.

The symptoms were the same: low charge, more oil, and no apparent leak. The technician again cleaned up the oil, but, this time, added fluorescent dye with the intent of returning with an ultraviolet light and black plastic tarp after a couple of days of operation.

A couple of days later, the ultraviolet light showed the dye signature of additional oil in the compressor compartment. The technician called yet another technician, Charlie, to get a second set of eyes on the situation. Charlie sat with the unit for a couple of hours. Finally, Charlie realized that, when the wind paused, he could hear a very faint noise that sounded like a leak in the condensing unit. Like a contortionist, he moved his head around the compressor compartment, trying to locate the source of the sound. Eventually, he realized the noise was coming from the condensing unit's control compartment, above the compressor section. He removed the cover, finding the expected electrical wires as well as more oil. (Keep in mind that no refrigerant piping is located in the control compartment.) The leak noise was a little louder, but still hard to pinpoint. The power was off, so Charlie started to soap the electrical compartment. Bubbles began to form at a wire nut.

Further investigation found a leak at a pressure switch. Somehow, the refrigerant did not escape at the switch, but instead traveled up the electrical wire (inside of the insulation) to the control compartment, where it escaped at the wire nut. The switch and connecting wire were replaced, and the problem was solved.

I hope I never again to have to find a refrigerant leak at a wire nut. This truly was a case of “Seeing is believing.”
Mike Gallagher, PE
Western Allied Corp.
Santa Fe Springs, Calif.

Have a “war story” to share? Send it to Executive Editor Scott Arnold at scott.arnold@penton.com.